Agent Incident RegistryAIR
v1 · updated 6 Oct 2026
Safety failure Realized harm

Air Canada held liable for its website chatbot's false bereavement-fare policy (Moffatt v. Air Canada, 2024 BCCRT 149)

AIR-2024-0008 AIR record · 2024-02-air-canada-chatbot-bereavement-fare
A real party was affected. This incident caused actual consequences, not a demonstrated capability.
From the supporting source
While a chatbot has an interactive component, it is still just a part of Air Canada's websitemccarthy.ca
The record
Disclosed
2024-02-14
Occurred
2022-11-11
Vendor
Air Canada
Product
Air Canada website support chatbot
Risk domain
Agent behaviour › Human manipulation
Also touches
Unsafe actuation
System type
Generative model
Who acted
Human acted on output (contested) — the chatbot stated a false policy and the passenger booked on it
Attack surface
Consumer assistant
Autonomy level
Assistant
Initial vector
Model hallucination
Impact type
Financial loss, Legal liability, False commitment
Confidence
High
Submitted by
EAI Internal
Crosswalk
OWASP Agentic Top 10 (2026)
ASI09 – Human-Agent Trust Exploitation
MITRE ATLAS
AML.T0054 – LLM Jailbreak
NIST AI RMF
MAP 5.1
EU AI Act
Recital 45
ISO/IEC
TR 24028; 42001; 23894
AIUC-1
C. Safety
Scope
agent misuse
Mitigation
Communicate model limitations; Require human oversight; Watermark AI content; Design for explicit consent.
Taxonomy
Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
one customer (CAD $650.88 awarded); precedent binding on operators of customer-facing chatbots in BC
Capability abused
policy_statement_on_behalf_of_vendorcustomer_facing_authoritypolicy_statement_on_behalf_of_operator
Defense & remediation
Recorded defense
correct policy published on a linked page the chatbot itself pointed to
Guardrail kinds
Unknown
Guardrail outcome
Silent
Fix shipped
Air Canada disabled the chatbot after the decision (reported)
Taxonomy & references
OWASP ASI03 Identity and Privilege AbuseOWASP ASI09 Human-Agent Trust Exploitationbinding_false_commitmentnegligent_misrepresentationhallucinated_policyoperator_liabilityagent_binds_principal
Curator notes
Boundary call, included deliberately: the chatbot emitted text rather than calling a tool, but it held delegated authority to state company policy and a tribunal made that statement legally binding on the vendor - the defining case for 'agent makes a commitment the principal must honour'. The primary decision (canlii / decisions.civilresolutionbc.ca) is not quoted here; used a law-firm analysis that quotes the decision verbatim. Decision date reported as 2024-02-14 (some sources say Feb 19). Also reported as: 'Moffatt v. Air Canada: airline held liable for its website chatbot's incorrect bereavement fare advice after arguing the chatbot was a separate legal entity'.
Cite as
Enkrypt AI. "Air Canada held liable for its website chatbot's false bereavement-fare policy (Moffatt v. Air Canada, 2024 BCCRT 149)". AIR, AIR-2024-0008. https://agentincident.vercel.app/i/AIR-2024-0008/ (accessed 2026-10-06).